What is page rank algorith?

[weegillis]

A very good illustration of the algorithm was posted by @deepsand in an earlier SEO89: thread:

PageRank is a measure of the probability that a page will be randomly requested. The mathematical function that describes it is ...

http://www.webproworld.com/webmaster...l=1#post624480

Because of the massive number of iterations involved, it is practically impossible to calculate one's own PageRank since one has really no real idea how many pages are linking, and how many links are in all those pages (combined).

We need also to bear in mind that in light of quality considerations, this algorithm is just a starting point. Brin/Page bring up this consideration in point #8, which has clearly been the path that Google has taken (to consider advertisers). PR may be calculated against the algorithm, but for SERPs it may have little or no bearing when all other factors are weighed in.

[C0ldf1re]

... If I did have the answer would I reveal it here or any other forum? Not a chance

Of course not! The problem comes when you finally figure out how to beat the Google algorithm. You then use your secret knowledge to get to the top of serps for, "Buy v14gr4 online", and the money starts rolling in. But your secrets are no longer secret -- SEO experts have descended on your site like a pack of wolves; your HTML is visible, and your backlinks are traceable. Just as your competitors copy your techniques and start to catch you in serps, you squeeze the last money from your system by selling an ebook: "Copy my system for being Google #1". Then everybody knows your techniques, and Google change their algorithm. So you start all over again...

[deepsand]

We can not explain about this in a single line, it has more and Thanks steve for providing wonderful resource.

To contrary, it is easily described in a single sentence, as here follows.

PageRank is a measure of the probability that a page will be randomly requested, the mathematical function that describes it being



where:

• where p_1, p_2, ..., p_N are the pages under consideration;
• M(p_i) is the set of pages that link to p_i;
• L(p_j) is the number of outbound links on page p_j;
• d is the damping factor; and,
• N is the total number of pages in the calculation matrix.

[ozsubasi]

divided by a dampening factor of 0.85 ?

Yes but quoting a figure without an explanation would not help someone to understand it. I couldn't find a way to do that so I left it out. I think that sometimes less is more, and the description I quoted was meant as a general one that perhaps some could more easily relate to, even if it wasn't complete.

[deepsand]

Yes but quoting a figure without an explanation would not help someone to understand it. I couldn't find a way to do that so I left it out.

Google has never divulged the value that it uses for the damping factor; and, it's not necessarily a fixed value across all time.

Either Page or Brin - don't remember which - made a statement that, at the time, it was estimated that there was an 85% probability that a user would tire of a site and move on to a URL chosen by means other than clicking on a link on the page presently being viewed. This would mean that (1-d)=0.85, and d=0.15.

[weegillis]

Google has never divulged the value that it uses for the damping factor; and, it's not necessarily a fixed value across all time.

Either Page or Brin - don't remember which - made a statement that, at the time, it was estimated that there was an 85% probability that a user would tire of a site and move on to a URL chosen by means other than clicking on a link on the page presently being viewed. This would mean that (1-d)=0.85, and d=0.15.

Which we can see, but we can also see that the damping factor is still applied 'en mass' and is still not taking into account true user interest beyond a mere assumption after all accounts. Which goes to say it had no business being there in the first place. Chalk up one flaw to this paper.

And with the very introduction of 'd', they invariably cannot give it a value of zero. This is a bias, not a damping factor, and arbitrary at that.

[deepsand]

Which we can see, but we can also see that the damping factor is still applied 'en mass' and is still not taking into account true user interest beyond a mere assumption after all accounts. Which goes to say it had no business being there in the first place. Chalk up one flaw to this paper.

PageRank is not a measure of a deterministic process, but of a stochastic (random) one. (There are other page rank functions developed outside Google which are different in form, but still stochastic in nature.)

And, the presence of the damping factor as here employed is the mathematically required way of determining probabilities re. a stochastic process with but two choices after each action.

The 2nd term in the function is the probability that one will randomly chose a new page via a link on the one presently being viewed; the 1st term, that he will chose by some other means. As these two choices are mutually exclusive, the sum of the probabilities of each choice must be equal to 1.

And with the very introduction of 'd', they invariably cannot give it a value of zero.

If a page has out-bound links, then the probability of their being followed cannot be zero.

This is a bias, not a damping factor, and arbitrary at that.

Properly speaking, (1-d) is the damping factor here. A high value causes reiteration of the function to converge more rapidly than a low value does. Hence the name "damping" or "decay" factor.

[blognama]

What is Google's page rank algorithm? What are the factors of it and how can we understand it?

PR could be understood clearly with the help of varieties of examples of different websites.There is not a set formula for PR (and if it is-then it won't be known so easily,so openly ) You can understand how the PR value works in a very general way and not like a mathematical formula.
By the way, writing quality and unique content is a big plus in increasing PR value and the reverse of it will work against the PR value.There could be more such similar points that can help you understand PR(in General). Don't go after a formula, make one for yourself

[williamc]

There is not a set formula for PR (and if it is-then it won't be known so easily,so openly )

Yes, there is. The patented formula is shown above.

By the way, writing quality and unique content is a big plus in increasing PR value

and the reverse of it will work against the PR value.

These are completely false.

[weegillis]

If a page has out-bound links, then the probability of their being followed cannot be zero.

Properly speaking, (1-d) is the damping factor here. A high value causes reiteration of the function to converge more rapidly than a low value does. Hence the name "damping" or "decay" factor.

That explains it. Was counting on someone setting me straight.